Impacts of the use of Nonnative Commercial Bumble Bees for
Pollinator Supplementation in Raspberry
Author(s): G. C. Lye, S. N. Jennings, J. L. Osborne, and D. Goulson
Source: Journal of Economic Entomology, 104(1):107-114. 2011.
Published By: Entomological Society of America
DOI: 10.1603/EC10092
URL: http://www.bioone.org/doi/full/10.1603/EC10092
BioOne (www.bioone.org) is an electronic aggregator of bioscience research content,
and the online home to over 160 journals and books published by not-for-profit societies,
associations, museums, institutions, and presses.
Your use of this PDF, the BioOne Web site, and all posted and associated content indicates
your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.
Usage of BioOne content is strictly limited to personal, educational, and non-commercial
use. Commercial inquiries or rights and permissions requests should be directed to the
individual publisher as copyright holder.
BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers,
academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.
FIELD AND FORAGE CROPS
Impacts of the Use of Nonnative Commercial Bumble Bees for
Pollinator Supplementation in Raspberry
G. C. LYE,1,2 S. N. JENNINGS,3 J. L. OSBORNE,4
AND
D. GOULSON1
J. Econ. Entomol. 104(1): 107Ð114 (2011); DOI: 10.1603/EC10092
ABSTRACT Evidence for pollinator declines has led to concern that inadequate pollination services
may limit crop yields. The global trade in commercial bumble bee (Bombus spp.) colonies provides
pollination services for both glasshouse and open-Þeld crops. For example, in the United Kingdom,
commercial colonies of nonnative subspecies of the bumble bee Bombus terrestris L. imported from
mainland Europe are widely used for the pollination of raspberries, Rubus idaeus L. The extent to
which these commercial colonies supplement the services provided by wild pollinators has not been
formally quantiÞed and the impact of commercial bumble bees on native bees visiting the crop is
unknown. Here, the impacts of allowing commercially available bumble bee colonies to forage on
raspberry canes are assessed in terms of the yield of marketable fruit produced and the pollinator
communities found foraging on raspberry ßowers. No differences were found in the abundance,
diversity, or composition of social bee species observed visiting raspberry ßowers when commercial
bumble bees were deployed compared with when they were absent. However, weight of marketable
raspberries produced increased when commercial bees were present, indicating that wild pollinator
services alone are inadequate for attaining maximum yields. The Þndings of the study suggests that
proportional yield increases associated with deployment of commercial colonies may be small, but that
nevertheless, investment in commercial colonies for raspberry pollination could produce very signiÞcant increases in net proÞt for the grower. Given potential environmental risks associated with the
importation of nonnative bumble bees, the development of alternative solutions to the pollination
deÞcit in raspberry crops in the United Kingdom may be beneÞcial.
KEY WORDS Rubus idaeus, Bombus terrestris, pollination limitation, Scotland, agriculture
A potential pollinator deÞcit for the production of
entomophilous crops is an increasing global concern
as a result of apparent declines in a range of pollinator
species worldwide (Allen-Wardell et al. 1998, SteffanDewenter et al. 2005). Although long-term trends in
crop production reveal no current global pollination
problem (Aizen et al. 2008), there is an increasing
dependence on animal-pollinated plants that is not
being met by increases in pollinator populations
(Aizen and Harder 2009). Bumble bees (Bombus spp.)
provide a superior pollination service for many ßowering crop plant species (Free and Williams 1976,
Stanghellini et al. 1997, Thomson and Goodell 2001),
but declines in these insects have been occurring
throughout their range (Kosior et al. 2007, Colla and
Packer 2008, Xie et al. 2008, Grixti et al. 2009). Bumble
bee declines in Europe have largely been attributed to
changes in land management associated with the rise
of intensive agriculture which has resulted in the re1 School of Biological and Environmental Sciences, University of
Stirling, Stirling FK9 4LA, United Kingdom.
2 Corresponding author, e-mail: [email protected].
3 MylneÞeld Research Services Ltd., Scottish Crop Research Institute, Invergowrie, Dundee DD2 5DA, United Kingdom.
4 Department of Plant and Invertebrate Ecology, Rothamsted Research, Harpenden, Hertfordshire AL5 2JQ, United Kingdom.
duction of ßoral abundance and diversity in the rural
environment (Goulson et al. 2008, Williams and Osborne 2009). In the United Kingdom at least, it seems
that those plant species that provide forage resources
for bumble bees have been disproportionately affected by these changes (Carvell et al. 2006). As a
result, bumble bee declines are particularly apparent
in the agricultural environment and evidence suggests
that rural areas now support lower densities of bumble
bee colonies that do urban areas (Goulson et al. 2002,
Osborne et al. 2008b).
In response to an apparent or perceived insufÞciency of natural pollinators, it has become increasingly common for farmers of certain entomophilous
crops to buy commercially available bumble bee colonies for Þeld use to supplement the pollination service provided to their crops. Currently, ⬇30 Ð 60,000
colonies per yr are imported into the United Kingdom
for the pollination of greenhouse or Þeld crop plants.
However, this practice may be associated with detrimental ecological consequences, including pathogen
spillover from commercial colonies to wild bumble
bee populations, competitive interactions between introduced bumble bees and local pollinators, and introgression of native and nonnative subspecies of B.
terrestris (Goulson et al. 2008).
0022-0493/11/0107Ð0114$04.00/0 䉷 2011 Entomological Society of America
108
JOURNAL OF ECONOMIC ENTOMOLOGY
Although the import of commercially reared
bumble bee colonies for Þeld crop pollination is a huge
industry, little research has been carried out to determine the need for pollinator supplementation for
these crops in the United Kingdom. Although growers
are unlikely to undertake this practice if yield loss by
pollinator limitation is not signiÞcant, the effectiveness of current populations of native bees for producing good yields of high-quality crop, and the capacity
of commercial colonies to supplement this have not
been independently quantiÞed. Soft fruit production
is a signiÞcant proportion of the market for commercially reared bumble bee colonies in the United Kingdom, and one such example is raspberry, Rubus
idaeus L. Raspberry production in Europe has been
estimated at 316,000 metric tons per yr (www.
fruitgateway.co.uk). Raspberry plants produce aggregate fruit consisting of several single drupelets, each of
which must be individually fertilized to develop. Commercial raspberry cultivars are self-compatible (Colbert and de Oliveira 1990, Willmer et al. 1994) and the
exclusion of pollinators is generally found to have no
negative effect on fruit set (Couston 1963, Szklanowska and Wienlarska 1993, Cane 2005 but see
Shanks 1969). This is because raspberry ßowers often
autopollinate as a result of contact between the stamens and the outermost ring of pistils (Cane 2005).
However, the innermost pistils are generally not fertilized in the absence of insect pollinators, and this
results in a terminal tuft of undeveloped drupelets
(Shanks 1969, Szklanowska and Wienlarska 1993) rendering the fruit unsuitable for marketing. Insect visitation to raspberry ßowers has been found to result in
increased drupelet number and fruit weight (Couston
1963, Shanks 1969, Chagnon et al. 1991, Szklanowska
and Wienlarska 1993, Cane 2005), and there is also
evidence that some raspberry cultivars may demonstrate metaxenia (an enhanced development of maternal tissues as a result of fertilization by pollens of
other varieties) giving rise to heavier fruit as a result
of pollen transfer among cultivars (Colbert and de
Oliveira 1990). Therefore, insect-mediated cross-pollination may provide economically important yield
increases for raspberry production.
Raspberry ßowers produce large quantities of nectar (Whitney 1984, Willmer et al. 1994) and are known
to be highly attractive to bees (Free 1968). Attributes
of this species are thought to be especially well suited
to promoting visitation by bumble bees (Whitney
1984), which are extremely effective pollinators of this
species due to their particular foraging behavior, morphology, and ability to forage in adverse environmental conditions (Gyan and Woodell 1987, Willmer et al.
1994). Many soft fruit growers in the United Kingdom
now buy commercially reared bumble bee colonies for
use in Þelds of raspberries to boost local pollinator
availability. However, because many of the bee species native to the United Kingdom are highly attracted
to raspberries (Willmer et al. 1994), it is possible that
pollinator limitation for this crop is low.
Here, the effect of the use of commercially available
bumble bee colonies on raspberry yields in central
Vol. 104, no. 1
Scotland is assessed. The effect of the presence of
these colonies on the natural pollinator communities
foraging on the raspberry ßowers also is examined to
look for evidence of competitive interactions between
commercial bumble bees and native social bees.
Materials and Methods
Fieldwork. Fieldwork was carried out at the Scottish Crop Research Institute (SCRI) in Invergowrie,
central Scotland. The experimental site consisted of
⬇0.5 ha of land planted with rows of raspberry canes
of mixed genotypes. Replicate plots consisted of Þve
canes of a single genotype planted in close proximity
to one another along a portion of a row. All plots were
uncovered during the Þrst three weeks of the study,
but open-ended polytunnels were erected over these
at the end of the third week of recording, following
normal commercial practice. Raspberry plants were
provided with water and nutrients through an irrigation system which allowed the precise delivery of
quantities of water and liquid feed calculated to maximize raspberry yields and ensure that these resources
do not limit fruit production. Four commercially
reared colonies of B. terrestris (Koppert Biological
Systems, Berkel en Rodenrijs, The Netherlands) were
placed at one end of the rows of raspberries, so that the
density of colonies was consistent with that recommended by the retailers (six to nine colonies per ha).
These colonies were opened or closed on alternate
weeks of the study such that during 1 wk the imported
bees were free to forage, but during the following
week, they were contained within their colonies. This
approach continued for 6 wk (three open, three
closed), commencing on the 24 May, starting with the
colonies open. During the weeks when the colonies
were closed, colonies were provided with sugar solution ad libitum and ⬇20 g of pollen per week (collected from honey bee hives by using pollen traps).
Access to sugar solution was maintained during periods when colonies were open because this is standard
practice by commercial growers when using colonies
for pollination of Þeld crops. When colonies were
closed, an entrance containing a one-way valve allowing trafÞc into the colony but not out remained open
to allow access to any foragers spending the night
away from the colony.
Fifty replicate plots of Þve plants (distributed across
three rows of canes) were selected for inclusion in the
study. These included twenty-nine different novel genotypes (developed as part of the SCRI Rubus breeding and genetic research program; Harrison et al.
1996) ranging in representation from one to Þve plots
each. (Each plot contained plants belonging to just
one of these genotypes). Plots were visited twice a
week at 1000 hours between 26 May and 3 July. During
each visit, a 100-m transect was walked at a slow and
constant pace along the length of each of the three
rows and the number of insects visiting ßowers were
recorded for each plot of Þve plants. Each transect
took ⬇15 min to complete. because the majority of
observations were of social bees (Apis mellifera L. or
February 2011
LYE ET AL.: RASPBERRY POLLINATION BY COMMERCIAL BUMBLE BEES
those belonging to the genus Bombus), only these
species were classiÞed to species level. Native Bombus
terrestris audax Harris, commercial Bombus terrestris
dalmatinus Dalla Torre, and bees belonging to the
Bombus lucorum L. complex (Murray et al. 2008) were
combined in a single group due to difÞculties in reliably distinguishing workers of these taxa in the Þeld.
After observations of insect visitation, the number of
receptive ßowers per plot of Þve plants was recorded,
and 10 randomly selected ßowers were marked with a
small twist of colored wire around the pedicel. Where
fewer than 10 ßowers were present, all ßowers were
marked in this way. The color of the wire was indicative of the day upon which the ßower was marked.
Ripe fruit present on raspberry canes was handpicked twice weekly between 1 July and 10 August.
After each picking, the weight of marketable fruit
produced from each plot was recorded. Wires also
were collected to allow assessment of the time taken
for ßowers to develop into fruit. The number of unfertilized ßowers and crumbly fruit (in which several
drupelets failed to develop giving rise to a deformed
fruit) marked with each wire color were recorded
allowing assessment of fruit quality between the two
treatments.
Weather data for the duration of the study period
were obtained from the UK Met OfÞce weather station in Leuchars, Fife.
Statistical Analysis. All analyses were carried out
using SPSS, version 16.0 (SPSS Inc., Chicago, IL). Five
plots were removed from the experimental site during
the course of the investigation as a result of poor plant
health and these were excluded from all analyses.
The number of social bee visits observed per raspberry ßower recorded was calculated for each recording day. In addition, the ratios of B. terrestris/B. lucorum visits to the sum of all other social bee visits and
a SimpsonÕs index of diversity for social bee visits also
were calculated. MannÐWhitney U tests were used to
compare these measures between days when imported colonies were free to forage (colonies open)
and days when the imported bees were contained
within their colonies (colonies closed).
For each genotype, the time taken for each tagged
ßower to develop into a ripe fruit was calculated and
a mean taken across all tagged fruit. This value was
taken to be representative of the approximate time
taken for all ßowers belonging to that genotype to
develop into fruit such that data collected during the
Þrst recording period could be paired with appropriate yield data. The proportion of total tagged ßowers
that failed to develop into marketable fruit either
because ßowers did not develop or because the number of drupelets that developed was low giving rise to
“crumbly” fruit was calculated per day of tagging, and
MannÐWhitney U tests were used to look for differences in these depending on colony status (open versus closed).
A linear mixed-effects model was built to assess the
effect of colony status on the weight of fruit produced
per Þve plant plot. The response variable was total
weight of ripe marketable fruit harvested per plot per
109
Table 1. Total number and percentage contribution of different insect taxa observed visiting raspberry flowers
Pollinator taxon
No.
visits
% total
visits
A. mellifera
B. hortorum
B. lapidarius
B. monticola
B. pascuorum
B. pratorum
B. terrestris/B. lucorum
B. bohemicus
Other (non-Bombus/Apis)
Total
102
1
16
43
73
299
1,427
1
20
1,982
5.1
0.1
0.8
2.2
3.7
15.1
72.0
0.1
1.0
100.0
time point (log transformed). The repeated measures
and nested elements of the experimental design were
incorporated into the model by specifying plot and
date as random effects (with plot included as the
subject variable). Fixed effects included in the initial
model were plant genotype, distance from the nearest
commercial colony, ßower number counted during
the initial recording period (log transformed) and
colony status during the initial recording period (open
or closed). An interaction effect between colony status and distance from the nearest colony and between
colony status and genotype also were tested for in this
study. Yield scores with a value of 0 were excluded
from the analysis because raspberries were only harvested when a signiÞcant number of fruit had become
ripe, thus observations of 0 did not represent yield of
0 g but simply the lack of raspberry harvesting on that
day. Calculating yield-to-ßower ratios exposed six obvious outliers. Because all outliers were associated
with low ßower counts, these outliers are likely to
have resulted from overestimation of yield per ßower
as a result of the inclusion of fruit derived from ßowers
from one or more of the previous ßower counts in the
corresponding yield measurement. Analyses were
conducted with and without these outliers. Because
removal of outliers did not affect the qualitative patterns observed results presented here exclude outlying data. The Þnal model was created by stepwise
removal of factors.
We used t-tests to compare daily averages of temperature and daily rainfall in weeks when colonies
were open versus closed.
Results
Many taxa were observed feeding on raspberry
ßowers, including several solitary bee and syrphid ßy
species. However, 99% of individuals observed belonged to A. mellifera or a Bombus species. At least
nine species of social bee were represented, but just
three species categories (B. terrestris/B. lucorum, B.
pratorum, and A. mellifera) constituted the majority of
observations (Table 1).
Abundance of social bee visits per raspberry ßower
was not affected by colony status (U ⫽ 13, P ⫽ 0.49)
nor was the ratio of B. terrestris/B. lucorum to other
social bee species (U ⫽ 13, P ⫽ 0.49). Social bee
110
JOURNAL OF ECONOMIC ENTOMOLOGY
Vol. 104, no. 1
Fig. 1. Mean proportion contribution of different species to observed social bee visits to raspberries per day when
commercially reared bumble bee colonies were open (free to forage) versus closed (foragers contained) ⫾ SE. B. terr/luc
refers to all individuals belonging to wild or commercial B. terrestris or any of the B. lucorum species complex.
diversity on raspberry ßowers (as measured by SimpsonÕs index of diversity) was the same regardless of
colony status (U ⫽ 14, P ⫽ 0.59). Overall, there was no
apparent effect of allowing commercially reared colonies to forage on the species composition of social
bees foraging on raspberry ßowers (Fig. 1).
Three thousand seven hundred and forty four tags
were reclaimed from ripe raspberry fruit. The average
time taken for tagged raspberry ßowers to develop
into ripe fruit varied among genotypes and ranged
from 34 to 41 d. Based on their means, genotypes were
assigned to one of three groups for which the mean
times from ßower to fruit were 35 d (SE ⫾0.22, N ⫽
493), 37 d (SE ⫾0.13, N ⫽ 1905), and 40 d (SE ⫾0.14,
N ⫽ 1346). Recording period pairs based on these data
were 34, 38, and 41 d (⫾1) apart, respectively. Failure
rate of marked fruit due to undeveloped ßowers or
crumbly fruit was low (1.7% marked ßowers), and
there was no difference in failure rates between proportions of failed fruit due to undeveloped ßowers or
proportions of crumbly fruit per recording period
when colonies were open or closed (U ⫽ 10.5, P ⫽ 0.24
and U ⫽ 9.5, P ⫽ 0.18, respectively).
As would be predicted, number of ßowers counted
in the initial recording period was strongly positively
correlated with weight of marketable fruit picked on
the corresponding date in the second recording period
(F ⫽ 594.34; df ⫽ 1, 309; P ⬍ 0.001), and yield of
marketable fruit also differed signiÞcantly among genotypes (F ⫽ 14.87; df ⫽ 28, 289; P ⬍ 0.001). When
other factors were taken into account, a greater weight
of marketable raspberries was picked in time periods
corresponding to weeks in which colonies were open
compared with weeks in which they were closed (F ⫽
6.75; df ⫽ 1, 336; P ⫽ 0.010). This was reßected by clear
differences in raw yield (when divided by corresponding ßower number) during the Þrst 4 wk of the study
but was not clear in the Þnal 2 wk (Fig. 2). When yield
was standardized by division by number of receptive
ßowers open during the corresponding recording period, the increase in average yield associated with
allowing the commercial bumble bee colonies to for-
Fig. 2. Average yield of raspberries per ßower counted in the corresponding recording period ⫾ SE.
February 2011
LYE ET AL.: RASPBERRY POLLINATION BY COMMERCIAL BUMBLE BEES
age was 27.93%. Distance of plots from colonies did not
affect the total yield of fruit (F ⫽ 2.13; df ⫽ 1, 320; P ⫽
0.145), nor was there any interaction effect between
distance of plots from colonies and colony status (F ⫽
0.97; df ⫽ 1, 282; P ⫽ 0.681) or genotype and colony
status (F ⫽ 1.43, df ⫽ 28, 226; P ⫽ 0.081).
Average weekly weather was found to be similar
throughout the duration of the study (data not
shown), and no signiÞcant difference was found in
daily averages of temperature or daily rainfall in weeks
when colonies were open versus weeks when they
were closed (t ⫽ 0.41; df ⫽ 41; P ⫽ 0.68 and t ⫽ 0.57,
df ⫽ 41, P ⫽ 0.57, respectively).
The parameter estimate for colony status obtained
from the model described above was used to calculate
the change in yield associated with the presence of
commercial colonies in the hypothetical situation in
which the log(yield) value per plot per harvesting
date in the absence of commercial colonies was 3.11
(the mean value when nests were closed) and all other
factors remain constant. In this case, raspberry yield
increases from 1,286.4 to 1,393.7 g per plot per harvesting date, an increase of 107 g or 8.3% (95% conÞdence intervals, 6.7%/10.3%). At the density at which
canes were planted at SCRI (⬇780 plots per ha), this
translates to a yield increase of 83.7 kg/ha per time
point. Because the average number of recording dates
for which yield was greater than zero for each plot was
nine, the total yield increase across the 6 wk of the
study would be 753.3 kg. At a market value of ⬇£6/kg
of Þrst class (for sale as fresh) raspberries in 2010, this
could translate into an increase in gross proÞt of
£4520/ha. The cost of providing bumble bee colonies
at the density used here is £336/ha (Koppert Biological Systems, in 2009) giving an increase in net proÞt
of £4,184/ha (95% CI, ⫺£1392/⫹£7841). If raspberries
are sold for pulp or for sale as frozen, market values
are much lower (⬇£0.7/kg and £2/kg, respectively),
reducing the estimated increase in net proÞt associated with the use of commercial bees to £191/ha
(95% CI, ⫺£459/⫹£618) or £1,170/ha (95% CI,
⫺£690/⫹£2,387), respectively.
Discussion
The observed yield increase in fruit produced from
ßowers receptive when commercial bumble bee colonies were allowed to forage compared with those
receptive when the colonies were contained suggests
that in the absence of commercial bumble bee colonies, pollination service does limit the yield of raspberry crops, at least at this site.
Despite the uniformity of the pollinator community
observed on the raspberry ßowers, the effect of commercial colony status on yield suggests that individuals
of the commercial colonies did visit the raspberry
ßowers during periods when they were allowed to
forage. It is therefore surprising that the relative contribution of B. terrestris/B. lucorum to other social bee
species observed foraging on raspberry ßowers did not
increase during these weeks. This observation could
either indicate that B. terrestris from commercial col-
111
onies were displacing native B. terrestris, B. lucorum
complex species, or both or that some aspect of the
experimental design resulted in undersampling of individuals from commercial colonies, for example, if
individuals from commercial colonies were active at
times of day other than that at which sampling took
place. The latter may be more likely because there was
no evidence for an impact of colony status on any
social bee species visiting raspberry, and it seems improbable that negative effects of competition from
commercial colonies should be restricted to wild B.
terrestris/lucorum.
The increase in raspberry yield associated with the
activity of commercially reared bumble bees in the
absence of any signiÞcant difference in the rate of
ßower failure suggests that yield differences were the
result of an increase in the weight of individual fruit
rather than in the number of marketable fruit produced. Drupelet number in fruit produced was not
counted, but it is likely that the increased weight of
raspberries that developed from ßowers pollinated
when commercial bees were free to forage was the
result of increased drupelet number due to more complete fertilization of ovules. However, because some
raspberry cultivars have been shown to demonstrate
metaxenic effects (Colbert and de Oliveira 1990),
greater rates of cross-pollination among genotypes
also may have contributed to this effect. Differences
in yield may represent a greater frequency of visits to
individual ßowers as a result of increased pollinator
numbers during periods when commercial colonies
were allowed to forage. However, it also is possible
that these differences could be partly attributable to
the differing characteristics of imported versus native
B. terrestris. Studies have shown that commercial colonies of B. terrestris demonstrate greater nectar-foraging efÞciency compared with laboratory reared
wild-caught B. terrestris audax (Ings et al. 2006), presumably the result of an increased rate of ßoral visitation per bee. Commercial B. terrestris also were
found to have a larger body size than native B. terrestris (Ings et al. 2006), so it may be that they transfer
more pollen grains per ßoral visit. Therefore, on an
individual basis, commercial bees might be expected
to provide a more effective pollination service than
native British bumble bees. In addition to this, it has
been shown that workers from commercial bumblebee colonies provided with nectar tend to focus on
pollen collection to a greater extent than those that are
not (Plowright et al. 1993). Because pollen-collecting
bees have been shown previously to transfer larger
quantities of pollen than nectar collectors (Goodell
and Thomson 2007), it is possible that the availability
of nectar within commercially reared colonies may
increase pollen transfer by commercial bees compared with wild bees simply as a result of differences
in foraging behavior.
Although differences in yield due to colony status
were clearly evident during the Þrst 4 wk of the study,
there was no observable yield difference during the
Þnal 2 wk (Fig. 2). This may reßect an increase in the
abundance of native bees as a result of seasonal colony
112
JOURNAL OF ECONOMIC ENTOMOLOGY
growth, although it also may be partly due to a deterioration in vigor of the commercial colonies toward
the end of the study (G.C.L., unpublished data).
A comparison of observations of bees visiting raspberry ßowers in this study (summer 2009) with those
reported in a study carried out at the same site (SCRI)
in the summers of 1990 and 1992 (Willmer et al. 1994)
suggests dramatic alterations in the local community
of social bees between these years. In the current
study, observations included at least seven bumble bee
species (B. terrestris/B. lucorum included as a single
species) as well as A. mellifera, whereas Willmer et al.
(1994) recorded just Þve bumble bee species (B. terrestris and B. lucorum recorded separately) and A.
mellifera. Regardless of commercial colony status
(open or closed) the B. terrestris/B. lucorum species
group contributed a much greater proportion to overall bee sightings in the current study (72%) than did
the B. terrestris and B. lucorum categories from
Willmer et al. (1994) (⬇30%), and B. pratorum also
was observed proportionately more commonly (15%
in this study versus ⬇8% by Willmer et al. (1994).
Conversely, A. mellifera and Bombus lapidarius L.
were far less well represented in the current study (5.1
and 0.8%, respectively) than in 1990 and 1992, in which
they made up ⬇34 and ⬇23% of observations. Willmer
et al. (1994) note that B. lapidarius was the commonest
bumble bee species observed at Invergowrie despite
its rarity in this region before the 1980s, but in the
current study this species was poorly represented and
has been relatively uncommon in central Scotland in
the summers of 2007 and 2008 (G.C.L., unpublished
data). The reduced number of A. mellifera observed
may indicate a similar drop in abundance of this species but because the positioning of domestic hives
during the studies is unknown, these differences may
simply reßect differences in local hive density. The
presence of Bombus monticola Smith visiting raspberry
ßowers in the current study is notable because this
species is usually associated with upland bog and heath
lands and also has shown range restrictions in the
United Kingdom over the past 60 yr (Goulson 2010).
The presence of this montane species at SCRI is surprising because Invergowrie is low lying, and the majority of the land surrounding this area comprises
urban space or arable agricultural land.
Willmer et al. (1994) also note differences in attributes between bee visitors observed in their study,
suggesting that some species might provide a more
effective pollination service for raspberries than others. B. lapidarius demonstrated particularly fast handling times of raspberry ßowers compared with other
Bombus species, and B. terrestris and B. lapidarius were
shown to transport more pollen grains between ßowers than Bombus pascuorum Scopoli and Bombus pratorum L., mainly as a result of their larger body sizes.
All species of Bombus were found to transport more
pollen grains than did A. mellifera. These Þndings
demonstrate that pollinator community composition
may be as important in determining pollination service
as pollinator abundance and also suggest that a decrease in proportion of visits by B. lapidarius could
Vol. 104, no. 1
potentially result in less effective pollination of raspberry ßowers.
Although this study demonstrates the value of commercial B. terrestris colonies for boosting raspberry
yields, there are ecological concerns associated with
the importation of these colonies into the United Kingdom (Goulson 2003, Ings et al. 2005, Ings et al. 2006).
Although B. terrestris has been demonstrated to pollinate raspberries very effectively (Willmer et al.
1994), studies have shown that other species also can
provide a high-quality pollination service for this plant
species. For example, the solitary bee species Osmia
aglaia Sandhouse is a superior pollinator of raspberry
ßowers in the United States (Cane 2005), and attributes of B. lapidarius suggest that it also can provide
a very effective pollination service for this species in
Europe (Willmer et al. 1994). At times of year when
native bumble bees are ßying, it is possible that measures targeted toward promoting populations of those
native bee species known to provide an efÞcient pollination service for raspberries may be able to achieve
similar increases in yield to bumble bee importation.
Such measures might include the planting of targeted
ßower strips or the provision of nesting sites for solitary bees. This approach could, in principle, reduce
the economic costs associated with the importation of
bumble bees and also eliminate any ecological risks
associated with the use of commercial colonies. The
inclusion of subsidies for such measures into appropriate agri-environment schemes also might be beneÞcial, providing an additional incentive for farmers to
consider these alternatives. Because the model presented in this study suggests that between 90 and 93%
of total yield is achieved in the absence of commercial
bumble bee colonies, the contribution by native pollinators is likely to be high, such that small increases
in their population sizes may be enough to negate the
effect of the commercial colonies. However, it must be
borne in mind that growers are able to extend the
fruiting period of raspberries with the use of hardy
varieties, polytunnels, and glasshouses such that timing of ßowering may not always correspond to pollinator phenology. In these situations, artiÞcially reared
colonies may be required to maximize yield in which
case, the use of native B. terrestris audax (newly available from Biobest Biological Systems in 2010) or the
rearing of B. lapidarius may provide an alternative
solution if the ecological risks of introducing nonnative subspecies are considered to be great.
In conclusion, this study suggests that yields of marketable raspberries can be increased by the presence
of commercially reared colonies of the bumble bee B.
terrestris. No evidence was found to suggest that commercial bees had any negative effects on species abundance or composition of other social bee pollinators,
although more thorough investigation would be required to satisfactorily exclude the possibility of competition between commercial bumble bees and wild
pollinator species. Although this study provides an
important insight into the use of commercially reared
bumble bee colonies for soft fruit pollination, much of
its value lies in highlighting the necessity for further
February 2011
LYE ET AL.: RASPBERRY POLLINATION BY COMMERCIAL BUMBLE BEES
study into a range of areas relating to this work. First,
the mechanisms behind yield differences remain unclear because there was no evidence that deployment
of commercial bumble bee colonies resulted in a
greater density of bees on the crop plant. Further work
is required to assess the propensity of commercially
reared bumble bees to forage on local crop plants and
any displacement of native bees of the same or closely
related species by individuals from imported colonies.
This could be achieved using a combination of massmarking individuals from commercially reared colonies (as in Osborne et al. 2008a) and analyzing pollen
loads coming into the colony to asses use of resources
(as in Whittington and Winston 2004). Secondly, the
pollination requirements of other soft fruit crops
should also be considered, because different plant
species will have differing pollination requirements
and may be differentially attractive to commercial or
native bees. Third, more research will be required to
elucidate the reality of the perceived environmental
risks associated with the importation of commercial
bumble bees into the United Kingdom. If these risks
are considered to be high, there is a need to develop
viable alternatives to the use of these colonies. This
could include both the development of methods of
boosting natural pollinator populations in soft fruit
growing regions, for example by the sowing of targeted
ßower mixes and/or provision of nesting sites, as well
as use of native bumble bee species for use at times
when natural pollinators are not active. In addition, a
comparison of observations made during this study
with those made at the same site in 1990 and 1992
demonstrates that we know little about the changing
community structure of pollinator populations.
Whether the observed differences reßect yearly ßuctuations or ongoing trends in species composition,
these differences are likely to have important implications for crop pollination and pollinator conservation into the future.
Acknowledgments
We thank Lynne Ferguson for Þeld assistance, Dave Warden for advice regarding raspberry growing, Mario VallejoMarṍn for statistical advice, and the Scottish Crop Research
Institute for the use of raspberry plots. We also thank Roy
Sanderson for comments on the manuscript. This project was
funded by the Biotechnology and Biological Sciences Research Council of the United Kingdom (project BB/
E000932/1).
References Cited
Aizen, M. A., L. A. Garibaldi, S. A. Cunningham, and A. M.
Klein. 2008. Long-term global trends in crop yield and
production reveal no current pollination shortage but
increasing pollinator dependency. Curr. Biol. 18: 1572Ð
1575.
Aizen, M. A., and L. D. Harder. 2009. The global stock of
domesticated honey bees is growing slower than agricultural demand for pollination. Curr. Biol. 19: 1Ð 4.
Allen-Wardell, G., P. Bernhardt, R. Bitner, A. Burquez, S.
Buchmann, J. Cane, P. Cox, V. Dalton, P. Feinsinger, M.
113
Ingram, et al. 1998. The potential consequences of pollinator declines on the conservation of biodiversity and
stability of food crop yields. Conserv. Biol. 12: 8 Ð17.
Cane, J. H. 2005. Pollination potential of the bee Osmia
aglaia for cultivated red raspberries and blackberries
(Rubus: Rosaceae). Hortscience 40: 1705Ð1708.
Carvell, C., D. B. Roy, S. M. Smart, R. F. Pywell, C. D.
Preston, and D. Goulson. 2006. Declines in forage availability for bumblebees at a national scale. Biol. Conserv.
132: 481Ð 489.
Chagnon, M., J. Gingras, and D. de Oliveira. 1991. Honey
bee (Hymenoptera: Apidae) foraging behavior and raspberry pollination. J. Econ. Entomol. 84: 457Ð 460.
Colbert, S., and D. de Oliveira. 1990. Inßuence of pollen
variety on raspberry (Rubus idaeus L.) development.
J. Hered. 81: 434 Ð 437.
Colla, S. R., and L. Packer. 2008. Evidence for decline in
eastern North American bumble bees (Hymenoptera:
Apidae), with special focus on Bombus affinis Cresson.
Biodivers. Conserv. 17: 1379 Ð1391.
Couston, R. 1963. The inßuence of insect pollination on
raspberries. Scott. Beekeeper 40: 196 Ð197.
Free, J. B. 1968. The foraging behavior of honeybees (Apis
mellifera) and bumblebees (Bombus spp.) on blackcurrant (Ribes nigrum), raspberry (Rubus idaeus) and strawberry (Fragaria ⫻ Ananassa) ßowers. J. Appl. Ecol. 5:
157Ð168.
Free, J. B., and I. H. Williams. 1976. Pollination as a factor
limiting the yield of Þeld beans (Vicia faba L.). J. Agric.
Sci. 87: 395Ð399.
Goodell, K., and J. D. Thomson. 2007. Inßuence of bee species (Hymenoptera: Apiformes) with contrasting behaviors on pollen movement in a mustard, Brassica rapa
(Brassicaceae) and the muskmelon Cucumis melo (Curcurbitaceae). Entomol. Gen. 29: 237Ð252.
Goulson, D. 2003. Effects of introduced bees on native ecosystems. Annu. Rev. Ecol. Evol. Syst. 34: 1Ð26.
Goulson, D. 2010. Bumblebees: behavior, ecology, and conservation. OUP, Oxford, United Kingdom.
Goulson, D., W.O.H. Hughes, L. C. Derwent., and J. C. Stout.
2002. Colony growth of the bumblebee, Bombus terrestris
in improved and conventional agricultural and suburban
habitats. Oecologia 130: 267Ð273.
Goulson, D., G. C. Lye, and B. Darvill. 2008. Decline and
conservation of bumble bees. Annu. Rev. Entomol. 53:
191Ð208.
Grixti, J. C., L. T. Wong, S. A. Cameron, and C. Favret. 2009.
Decline of bumble bees (Bombus) in the North American
Midwest. Biol. Conserv. 142: 75Ð 84.
Gyan, K. Y., and S.R.J. Woodell. 1987. Analysis of insect
pollen loads and pollination efÞciency of some common
visitors of four species of woody Rosaceae. Funct. Ecol.
1: 269 Ð274.
Harrison, R. E., R. J. McNicol, and S. Jennings. 1996. Rubus
breeding and genetic research, pp 63Ð 68. SCRI Annual
Report 1996. Scottish Crop Research Institute, Invergowrie, Scotland.
Ings, T. C., N. E. Raine, and L. Chittka. 2005. Mating preference of commercially imported bumblebees (Bombus
terrestris) in Britain (Hymenoptera: Apidae). Entomol.
Gen. 28: 233Ð238.
Ings, T. C., N. L. Wards, and L. Chittka. 2006. Can commercially imported bumble bees out-compete their native conspeciÞcs? J. Appl. Ecol. 43: 940 Ð948.
Kosior, A., W. Celary, P. Oloejniczak, J. Fijal, W. Król, W.
Solarz, and P. Plonka. 2007. The decline of the bumble
bees and cuckoo bees (Hymenoptera: Apidae: Bombini)
of western and central Europe. Oryx 41: 79 Ð 88.
114
JOURNAL OF ECONOMIC ENTOMOLOGY
Murray, T. E., U. Fitzpatrick, M.J.F. Brown, and R. J. Paxton.
2008. Cryptic species diversity in a widespread bumble
bee complex revealed using mitochondrial DNA RFLPs.
Conserv. Genet. 9: 653Ð 666.
Osborne, J. L., A. P. Martin, N. L. Carreck, J. L. Swain, M. E.
Knight, D. Goulson, R. J. Hale, and R. A. Sanderson.
2008a. Bumblebee ßight distances in relation to the forage landscape. J. Anim. Ecol. 77: 406 Ð 415.
Osborne, J. L., A. P. Martin, C. R. Shortall, A. D. Todd, D.
Goulson, M. E. Knight, R. J. Hale, and R. A. Sanderson.
2008b. Quantifying and comparing bumblebee nest densities in gardens and countryside habitats. J. Appl. Ecol.
45: 784 Ð792.
Plowright, R. C., J. D. Thomson, L. P. Lefkovitch, and C.M.S.
Plowright. 1993. An experimental study of the effect of
colony resource level manipulation on foraging for pollen
by worker bumble bees (Hymenoptera: Apidae). Can. J.
Zool. 71: 1393Ð1396.
Shanks, C.H.J. 1969. Pollination of raspberries by honeybees. J. Apicult. Res. 8: 19 Ð21.
Stanghellini, M. S., J. T. Ambrose, and J. R. Schultheis. 1997.
The effects of honey bee and bumble bee pollination on
fruit set and abortion of cucumber and watermelon. Am.
Bee J. 137: 386 Ð391.
Steffan-Dewenter, I., S. G. Potts, and L. Packer. 2005. Pollinator diversity and crop pollination services are at risk.
Trends Ecol. Evol. 20: 651Ð 652.
Szklanowska, K., and J. Wienlarska. 1993. The effect on
yield by the bees visits on raspberry ßowers (Rubus
Vol. 104, no. 1
idaeus L.) the successive ramiÞcations in inßorescence. In
K. Smolarz, and K. J. Zmarlicki (eds.), Sixth International
Symposium on Rubus and Ribes. Acta Hortic. 352: 231Ð
235.
Thomson, J. D., and K. Goodell. 2001. Pollen removal and
deposition by honeybee and bumblebee visitors to apple
and almond ßowers. J. Appl. Ecol. 38: 1032Ð1044.
Whitney, G. G. 1984. The reproductive biology of raspberries and plant-pollinator community structure. Am. J. Bot.
71: 887Ð 894.
Whittington, R., and M. L. Winston. 2004. Comparison and
examination of Bombus occidentalis and Bombus impatiens (Hymenoptera: Apidae) in tomato greenhouses. J.
Econ. Entomol. 97: 1384 Ð1389.
Williams, P. H., and J. L. Osborne. 2009. Bumblebee vulnerability and conservation world-wide. Apidologie 40:
367Ð387.
Willmer, P. G., A.A.M. Bataw, and J. P. Hughes. 1994. The
superiority of bumblebees to honeybees as pollinators:
insect visitors to raspberry ßowers. Ecol. Entomol. 19:
271Ð284.
Xie, Z., P. H. Williams, and Y. Tang. 2008. The effect of
grazing on bumblebees in the high rangelands of the
eastern Tibetan Plateau of Sichuan. J. Insect Conserv. 12:
695Ð703.
Received 11 March 2010; accepted 1 September 2010.